Docs · Concepts
How Ciela makes decisions.
Mira, Eli, Vance, and Atlas do the sending. Ciela does the thinking. She's the chat layer above the agents, the one who reads what Mira's getting on LinkedIn, what Eli's seeing in the inbox, what Vance heard on the last three calls, and tells you what to actually do next. This page is the honest account of where her answers come from.
What she is, what she isn't
- Role
- Chief of staff
- Lives at
- /dashboard/ciela
- Sends anything?
- No, she advises and drafts, the agents send
- Reads from
- Your real campaigns, contacts, replies, account state
- Grounded in
- Agency playbook knowledge base + live agent data
- Trust model
- Tells you when she's confident, tells you when she isn't
The three sources she reads from
- Your live data. Active campaigns, recent sends, replies, accept rates, booked meetings, account health. Pulled fresh on every chat turn so the answer reflects the state of your business right now, not last week.
- An agency playbook. A curated knowledge base of the moves that actually work for AI automation agencies, niche selection, pricing, the cold-call demo loop, retention. Ciela reaches into it for tactics, not lectures, and grounds them in your specific situation.
- Her chat with you. Not a memory across sessions, the in-conversation context, what you've told her this turn, what she said last turn, what you reacted to. She doesn't pretend to remember things you haven't told her yet.
How she gets from data to advice
- She looks at what you actually asked, not the wrapper around it. "Why is reply rate flat" gets a different read than "what should I do today", same data, different framing.
- She pulls the slice of your live data that's relevant. If the question is about email, she looks at Eli's last 30 days. If it's about whether to call a specific list, she looks at the list and Vance's recent work.
- She cross-references with the playbook. If the data says reply rate is 0.4% and the playbook says solid cold email lands between 1 and 3%, she'll tell you the diagnosis (subject, targeting, deliverability), not just the number.
- She gives you one move, not five. Lists with five options sound thorough, they're actually a way of saying "I don't know which one matters", Ciela doesn't do that. She picks one and tells you why.
What she will not do
- Won't fabricate metrics. If she doesn't have the data point you're asking about, she says so. She'd rather show you which surface to check than invent a number.
- Won't push send. Anything that goes to a real prospect comes from one of the four agents, never from Ciela. She'll draft, she'll plan, she'll review, she won't dispatch.
- Won't pretend she's certain when she isn't. On a question with a real tradeoff she'll lay the tradeoff out and say which side she'd lean and why, not pick blindly.
Where to talk to her
Open /dashboard/ciela and ask her anything about your business. The chat is the surface, there's no separate console, no widget to install. Most users open her at the start of the day to ask "what should I do today" and again when something interesting happens (a new reply, a flat metric, a list that just dropped).
Keep reading
Need help?
Ciela can walk you through any step. Or write to support@ciela.ai and we'll jump on the connection with you.